(PDF) Cateruca Processing Iron Ore

On the concentration of Cateruca iron ore Da concentração do minério de ferro de Cateruca Sobre la concentración de la mena de hierro de Cateruca DOI: 10.54021/seesv6n1-024 Originals received: 01/10/2025 Acceptance for publication: 02/03/2025 Henrique Kiaku Simão Master of Science in Mineral Engineering (PPGEM — UFOP) Institution: Tosyali Iron & Steel Angola S.A. Address: Techamutete, Huíla, Angola E-mail:

José Aurélio Medeiros da Luz Doctor in Mining Engineering (PPGEM — EEUFMG) Institution: Universidade Federal do Ouro Preto Address: Ouro Preto, Minas Gerais, Brazil E-mail:

Pedro Henrique Neuppmann Master of Science in Mineral Engineering (PPGEM — UFOP) Institution: Mosaic Fertilizantes do Brasil Ltda. Address: Catalão, Goiás, Brazil E-mail:

ABSTRACT A processing route for an iron ore sample from the Cateruca deposit in the Jamba mining district, Huíla Province, Angola, was studied. The ore contained 31 % Fe, and the main gangue mineral was quartz. Thermal analysis, X-ray diffractometry, and Mössbauer spectroscopy revealed the following iron mineral distribution: 90 % as hematite, 8 % as magnetite, and 2 % as goethite. Jigging was studied using a Denver mineral jig, keeping the hutch water flow rate and particle size constant while varying the ragging thickness. A Humphrey’s spiral was also tested by varying the solid content and, consequently, the feed flow rate, with the best result achieved at a solid content of 20 %. High-gradient magnetic separation was carried out as well, yielding good results under several magnetic fields. Reverse froth flotation, using amine as a quartz collector and starch as a depressant for iron-bearing minerals, resulted in a rougher concentrate with an iron grade of 64.19 % and a recovery of 93.5 %. The best depressant was corn starch. Direct flotation achieved an iron grade of 57.34 % in the concentrate during the rougher stage, with an iron recovery of 67.47 %. Sodium oleate was used as a collector for iron oxides, while ethanoic acid and water glass served as gangue depressants, and kerosene acted as a chain extender (collection booster). Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 1 Keywords: Iron ore. Jigging. Spiral. Flotation. Process development. RESUMO Uma rota de processamento para uma amostra de minério do depósito de Cateruca, no distrito mineiro de Jamba, província da Huíla, Angola, foi estudada. O minério continha 31 % de Fe, e o principal mineral de ganga era o quartzo. A termogravimetria, a difratometria de raios X e a espectroscopia Mössbauer revelaram a seguinte distribuição de minerais de ferro: 90 % como hematita, 8 % como magnetita e 2 % como goethita. A jigagem foi estudada com um jigue Denver, mantendo constantes a vazão de água de arca e o tamanho das partículas, enquanto se variou a espessura da cama. Um concentrador helicoidal (“espiral”) Humphrey também foi testada, variando-se a concentração de sólidos e, consequentemente, a vazão de alimentação, com o melhor resultado obtido sob concentração mássica de sólidos de 20 %. A separação magnética a úmido de alto gradiente apresentou bons resultados, sob diversas intensidades de campo magnético. A flotação reversa, utilizando amina como coletor de quartzo e amido como depressor dos minerais portadores de ferro, resultou em um concentrado bruto com teor de ferro de 64,19 % e recuperação de 93,5 %. O melhor depressor foi o amido de milho. A flotação aniônica direta alcançou teor de ferro de 57,34 % no concentrado da etapa de desbaste, com recuperação de ferro de 67,47 %. Oleato de sódio foi utilizado como coletor de óxidos de ferro, enquanto ácido acético e silicato de sódio como depressores de ganga, sendo querosene empregado como extensor de cadeia (reforçador de coleta). Palavras-chave: Minério de ferro. Jigagem. Concentrador Helicoidal. Flotação. Desenvolvimento de processos. RESUMEN Se estudió una ruta de procesamiento para una muestra de mena del yacimiento de Cateruca, en el distrito minero de Jamba, provincia de Huíla, Angola. La mena contenía un 31 % de Fe, y el principal mineral de ganga era el cuarzo. La termogravimetría, la difracción de rayos X y la espectroscopía Mössbauer revelaron la siguiente distribución de menas: 90 % como hematita, 8 % como magnetita y 2 % como goethita. La concentración por jig se estudió utilizando un jig Denver, manteniendo constantes el flujo de agua de arca y el tamaño de partícula, mientras se varió el espesor de la cama. También se probó un concentrador helicoidal ("espiral") Humphrey, variando la concentración de sólidos y, en consecuencia, el flujo de alimentación, obteniéndose el mejor resultado con una concentración másica de sólidos del 20 %. La separación magnética en húmedo de alto gradiente presentó buenos resultados bajo diversas intensidades de campo magnético. La flotación inversa, utilizando amina como colector de cuarzo y almidón como depresor de las menas, resultó en un concentrado bruto con una ley de hierro del 64,19 % y una recuperación del 93,5 %. El mejor depresor fue el almidón de maíz. La flotación aniónica directa alcanzó una ley de hierro del 57,34 % en el concentrado de la etapa de desbaste, con una recuperación de hierro del 67,47 %. Se utilizó oleato de sodio como colector de óxidos de hierro, mientras que el ácido acético y el silicato de sodio actuaron como depresores de ganga, empleándose queroseno como extensor de cadena (reforzador de recolección). Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 2 Palabras clave: Mena de hierro. Criba hidráulica. Desarrollo de procesos. Espiral. Flotación. 1 INTRODUCTION The banded iron formation is part of a volcanic-sedimentary sequence known as the Jamba Group, which is older than 2.5 billion years. The overall oxide assemblage of the Cateruca deposit closely resembles that of most Archean and Proterozoic iron formations. Many rare earth element patterns exhibit a small negative cerium anomaly and a positive europium anomaly, which is commonly associated with hydrothermal enrichment processes in banded iron formations. The Cateruca deposit is often classified as a banded hematite quartzite (BHQ) formation subtype (Sinués et al., 2011). This study focuses on analyzing the concentration characteristics of iron ore from the Cateruca deposit. It aims to comprehensively characterize the ore through particle size analysis, chemical analysis, X-ray diffractometry, and Mössbauer spectroscopy. This work has also evaluated concentration methods, encompassing jigging, spiral concentration, high gradient magnetic separation, and froth flotation, in order to determine their effectiveness in improving iron grade and recovery. Ultimately, this research will provide insights into optimal processing routes for Cateruca iron ore and its economic potential. 2 THEORETICAL FRAMEWORK Industrial sorting operations depend on the degree of liberation and particle size distribution of the ore constituents. There must also be sufficient contrast in one or more physical properties among these components to facilitate the selection of the most appropriate concentration method. In this type of ore, oxides and hydroxides are the most common categories of commercial interest among various iron-bearing minerals (Silva, 1999). When developing a mineral processing route, it is the rule to start with a technological characterization, which, both clarifies the intrinsic properties of the Cateruca´s mineral assemblage (and its textural interrelationships), and studies Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 3 the ore's response to concentration methods. Naturally, it starts with simpler sorting operations that often result in a higher internal rate of return on capital and less environmental impact. Thus, on a bench scale, where costs are relatively low, carrying out tests with gravitational methods is common practice, even if the degree of liberation of the mineralogical species does not recommend them (at least their single application). It is even common, in industrial applications, for (at least) a portion of complex ores to be directed to gravity circuits that require, in theory, higher values of the Taggart concentration criterion (Taggart, 1945) with the correction advocated by Burt (1984). The features of liberation among the mineralogical constituents of BIF-type iron ores are, not infrequently, misleading. They depend on several factors inherent to the local historical geology such as the intensity of tectonic phenomena, weathering, hydrothermalism, and remineralization. Therefore, simpler sorting methods should not be neglected a priori, even though they commonly result in greater difficulty in processing fines. In line with this reasoning, many techniques of iron ore beneficiation have been investigated in this work aiming at minimizing gangue content in concentrate with higher iron recoveries. Iron ore processing, not rarely, involves a gravity or magnetic pre-concentration step and subsequent concentration by froth flotation. It should be clear that this is a green field study, of technological characterization; being, therefore, an evaluation of the amenability of the ore (strictly speaking, of representative samples of the deposit) to the typical processes of its enrichment. If, let's say, a commercial concentrate with over 62 % iron content and about 11 % gangue is adopted as a target (values consistent with industrial practice for poor ores), obviously a route with rougher, cleaner, scavenger stages (in their multiple variations) should be adopted. This engineering step of increasing the performance of sequential operations (that is to say: establishing the topology and technological characteristics of the selected equipment) was beyond the scope of this article. Helix concentrators (called “spirals”) and jigs are extensively used as preconcentration step in countries like Australia, Brazil, India and South Africa. Jigging of iron ores has been practiced in these countries for many years due to relatively easy separation of minerals by this technique and capacity of Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 4 generating the correct stroke amplitudes to teeter the bed of heavy ore (Dieudonné, Jonkers and Loveday, 2006) Just to cite another example of gravity concentration, a jigging route was employed by Prakash and coworkers (2007) to treat an Indian iron ore. The results obtained were: iron concentrate content of 63.7 %; iron tailings content of 47.1 %, iron feed content of 59.4 %; recovery of 78.6 % (Das et al., 2007). Previously, some authors have achieved good results with jigging tests using African iron ores. The ore from Itakpe mine (Nigeria) contains mainly hematite, magnetite and quartz. Effective concentration of Itakpe iron ore by jigging operations proved to be feasible with an iron feed content of 36.0 % (Olubambi and Potgieter, 2005). In turn, jigging concentrates with content over 90.0 % magnetite were viably obtained by reprocessing typical flotation tailings from intrusive carbonatite (phosphate rock) processing (Silva et al., 2016). As iron ore jigging is concerned, several variables have already been studied, such as feed particle size and the type of ragging, as well as its thickness. Ragging thickness usually is adopted as three times the feed top size (Aplan, 2003). Particle size distribution has been considered the main factor on iron recovery. As a matter of fact, although usually jigs perform better with coarse particle sizes, sometimes feeds having a top size of 3.35 mm do result in lower iron recoveries, when compared with grain size below 2.36 mm (São José, Barcelos and Pereira, 2017). According to São José, Barcelos and Pereira (2017), ragging type has proved to be indifferent when hematite pebbles were used instead of steel balls. Finally, the ragging bed thickness has influenced directly the iron recovery. As expected, low ragging mass has caused higher denser phase recoveries (São José, Barcelos and Pereira, 2017). As a rule, interlocking of the ragging particles is avoided by the use of preferably rounded or sub-angular particles, especially when the ragging particle size is larger than 3 mm. For this reason, indurated iron ore pellets are occasionally used as ragging material. Helix concentrators (“spirals”) present better performance when processing sinter feed fines (0.15 mm to 1.00 mm size range) comparing to pellet feed (Arenare, Araujo, Viana and Rodrigues, 2008). On the other hand, magnetic separation and flotation techniques are among the most used techniques for Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 5 industrial beneficiation of iron ore. Although these methods have good efficiency, as a rule, they tend to produce iron concentrate with high amount of quartz in case of fine processing. Fines do not float very well in conventional flotation circuits because they, on the one hand, are very prone to the hydrodynamic drag (linked to the low particle momentum) and, on the other hand, do require higher collector dosages and longer flotation times (Sivamohan, 1990). As for magnetic concentration, there are many examples of adoption of this concentration technique for iron ore dressing (Silva, 2016; Luz, 2015). Just to cite an example, experiments were conducted by Dworzanowski (2012) to quantify the differences in magnetic separation performance with decrease in particle size. The degree of liberation is very important for magnetic separation (as, moreover, in any physical and physicochemical sorting operation), since — in this case — locked particles tend to have similar magnetic susceptibility to those magnetic parcels and can be trapped inside the grooved matrixes. Several developments were made to magnetic separation equipment to improve their efficiency. The SLon vertically pulsating high gradient magnetic separator (VPHGMS) system is the latest one that consists of a vertical carousel (i. e.: horizontal rotating axis) with rod matrix system and pulsation mechanism. It enables the achievement of higher capacities and improved separation performance and is more suitable to ultrafine particles or slimes (Dobbins, Dunn and Sherell, 2009). Brazilian companies are performing successful tests with this equipment on their iron ores to prove its efficiency. In this study, the focus was on Cateruca’s iron ore amenability concerning diverse methods of concentration (gravity separation, magnetic separation and froth flotation). This investigation constitutes a pioneering step, in order to ultimately subsidize further process development for the ore in question. 3 EXPERIMENTAL PROCEDURES 3.1 SAMPLE SPECIFICATIONS AND ITS PREPARATION Ore samples have been collected based on the geological maps from a study made by McKee (1974). About 200 kg of representative sample (boulders Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 6 and lumps) from Cateruca iron ore were collected and brought to the laboratory of mineral processing in Federal University of Ouro Preto (Brazil). Figure 1 illustrates the features of sample, displaying the zoning between layers of iron oxides and quartz. The picture was taken on a 7 mm aperture screen, in order to provide a scalable dimension. Figure 1. Sample of Cateruca banded iron ore (screen’s mesh aperture is 7 mm) Source: Prepared by the authors. Sample preparation was done as follows: firstly, the blocks were broken with a sledgehammer up to a suitable size for crushing. Then the fragments were processed in a jaw crusher (single toggle jaw crusher with feed area of 0.13 m x 0.10 m). The crushed sample was blended, homogenized, and then many representative subsamples were drawn by conning and quartering method for mineralogical and chemical characterization, and beneficiation studies, which included gravity, magnetic and flotation separation techniques. After the mentioned comminution stages (in closed circuit with 25 mm as mesh of crushing) the crushed sample size distribution fitted well to the Harris distribution given by following Equation 1 as a function of top size (xmax). 𝑥 𝑎 𝑏 ) ] 𝑌 = 1 − [1 − ( 𝑥𝑚𝑎𝑥 Where: (1) Y = cumulative passing fraction through the aperture x [mm]; Regression parameters: a = 0.918; b = 0.696; and xmax = 30.6 mm (sample theoretical top size). Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 7 The coefficient of determination (square root of regression coefficient) of the above statistical equation was R2 = 0.99324. The regression parameters were estimated by employing a non-linear algorithm from EasyPlot© software package. This algorithm uses a Marquardt– Levenberg filter, which — unlike the descending simplex search algorithm — allows estimating the uncertainty associated with the regression values (Karon, 2014). 3.2 SAMPLE CHARACTERIZATION To find out the distribution of iron grade, a representative crushed sample was sequentially crushed again in jaw crusher, roll crusher and, finally, in disc mill. The resulting sample was comminuted below 1.2 mm. Another representative sample was reground below 100 µm (by iron carbide rings in an orbital pan mill) and it was taken to diffractometric study, using an X-ray Shimadzu diffractometer (model XRD-6000), employing a monochromatic Cu-Kα (λ = 0.1542 nm) radiation, to discover the major mineral composition of the sample. Chemical analyses were performed to assess the contents of the chemical elements in head sample. The technique of inductively coupled plasma-atomic emission spectroscopy (ICP-AES) was used. The determination of iron assay from the ore’s different size fractions and from the Denver mineral jig products were done using a calibration curve from pycnometry, employing a helium pycnometer (model Ultrapyc 1200e V.4.00 from Quantachrome). Regarding the accuracy of this latter analytical procedure is appropriate to quote the work of Couto, Braga and França (2012) who have estimated the iron content in ore samples by gas pycnometry, showing that those estimated figures were very close to the corresponding ones obtained from X-ray fluorescence chemical analysis, resulting mean absolute deviation of only 1.2 %. Additionally in this present work, a study about this representativeness was made in order to adjust data and confirm previous estimation. In turn, the iron grade of samples from magnetic separation, pilot scale jig and flotation tests were done by the conventional chemical titration method (dichromatographic titration). Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 8 The investigation of the oxidation state of iron was done by Mössbauer spectrometry which helped to quantify the iron-bearing minerals in the sample, as used by Govaert et al. (1976). In addition, the thermogravimetric tests and differential thermal analysis were carried out on an apparatus DuPont SDT2960 STD V3.0F, with temperature range from 293 K to 1273 K (20 °C to 1000 °C), under heating rate of 0.333 K/s (20 oC/min) in air atmosphere. Additionally, the conventional Bond work index (conceptually the specific energy required to comminute a particle of infinite size to its progeny with 80% passing 100 micrometers) was estimated as 27.44 kJ/kg (6.915 kWh per short ton). 3.3 METALLURGICAL TESTWORK As far as the experimental procedure is concerned, it should be noted that the operational parameters chosen for the tests described below were within the typical range of similar processes. As a matter of fact, even due to the brevity required for a technical article, those prospective experiments that guided the choice of the values employed for the operating parameters have not been reported here. 3.3.1 Jig concentration Jigging is still routinely employed in the industrial production of sinter feed, despite its alleged high consumption of process water, which is easily solved with decantation and recirculation of the water to the jigging stage. It should be kept in mind these unit operations here performed only represent the rougher stage (in the detailed flowsheet phase, circuitry topology and equipment sizing should be undertaken, of course). Two different jigging routes were tested in this study. The first jigging experimental campaign took place in the Mineral Processing Laboratory of School of Mines at Federal University of Ouro Preto, using a Denver mineral jig (Figure 2). Jig’s piston stroke was kept constant at 8.0 mm and its frequency was 8.0 hertz (480 strokes/min). Aliquots of 0.700 kg of the ore were the feed for the bench scale jigging Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 9 operation. Monosized steel balls (average diameter of 5 mm) were spread to form a layer on the screen of the mineral jig as a bedding material (“ragging”) with varying thickness. The sample was fed manually, and water was added to the previous deaerated hutch through a controller valve. Each experiment was carried out for 5 minutes at a constant feed rate. The sample was collected after allowing the jig to stabilize for a lapse of 3.0 minutes (actually the hutch concentrate was somewhat contaminated during the first minute of operation). For this bench-scale jigging tests, different sized material was prepared by stage crushing in a laboratory roll crusher and disc mill, followed by screening at the desired size. The effect of ragging thickness on jigging was studied in a bench scale using the size fraction between 0.6 mm and 1.0 mm. In addition to this first campaign, jigging experiments were done using the size fraction between 1.0 mm and 3.0 mm but employing a manual cylindrical jig (0.50 m in diameter) in semi batch operation, without ragging. Figure 2. Denver jig with chamber effective area equal to 32 mm x 50 mm (0.0016 m2) Source: Prepared by the authors. For the sake of brevity, as jigging is concerned, particle size effect was not addressed in this paper, only the Influence of ragging thickness. The experiment using the Nomos’s manual cylindrical jig (without ragging) was carried out with water flow rate of about 0.300 L/s (resulting in a jigging chamber water surface velocity: Jf = 0.0015 m/s) and pulsation frequency of 3.69 Hz (221 strokes per minute). Sample of 11.0 kg was used in each batch operation of the semi pilot-scale Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 10 jig with hutch water being continuously added (44 mL/s) through a controller valve (actually becoming a semi-batch operation for this reason). Separation efficiencies were calculated for experiments with three ragging thicknesses. This parameter was calculated based on the iron content in feed, concentrate and tailings. The maximum iron content obtainable (ultimate content) was calculated according to the stoichiometric reasoning for hematite, considering the molar mass of each chemical element of the mineral (resulting value of 69.94 % Fe). The separation efficiency (E) was calculated as follows (Equation 2): 𝐸= 𝐶 × 𝑚 × (𝑐 − 𝑓) 𝐶 × 69.94 % × (𝑐 − 𝑓) = (𝑚 − 𝑓) × 𝑓 (69.94 % − 𝑓) × 𝑓 (2) Where: C = mass recovery (fraction of the total feed weight that reports to the concentrate) [–]; f = feed assay [–]; c = concentrate assay [–]; m = theoretically maximum iron content obtainable (0.6994) [–]. 3.3.2 Humphrey´s spiral concentration A representative sample of 55.0 kilograms was used for Humphrey´s spiral (helix) concentrator tests. The sample was recrushed in a roll crusher and in a disc mill in order to get the desired size range (between 0.054 mm and 1.000 mm). The Humphrey´s spiral test rig which was used had a 5-turn helicoid of 600 mm outer diameter (20 mm thickness cast iron wall). The effects of feed rate and percentage of solid content were analyzed. The procedure was as follows: after getting the sample totally comminuted, it was homogenized and quartered, and then an amount of dry sample was taken to prepare the slurry with three mass solid concentrations (20 %, 25 %, and 30 %). Water was added to the pump sump in the required amount. A recirculation slurry pump was turned on and 20.0 kg of the solid sample was added slowly, thereby preventing the solid settling at the sump bottom and pipe clogging. Once the slurry was completely constituted, the feed pump was turned on. The concentrate and tailing streams were collected after about 10 minutes of operation in closed circuit (in order to reach the steady state). After each sampling increment cut, the same amount of slurry removed was added to feed sump, so that the same solid Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 11 concentration could be kept (makeup slurry). All the products thus obtained were dried, weighed and analyzed. 3.3.3 High gradient wet magnetic separation In these experiments, two samples of 7.0 kg each were ground in a disc mill, one of them was ground below 1.00 mm size, and another one below 0.15 mm. The experiments were performed at Gaustec Magnetismo lab (a Belo Horizonte based company). The equipment used was a high gradient wet carousel-type magnetic separator (model Minimag®) with stainless steel matrix of 5.0 mm gap (slot) for the rougher stage and a gap of 2.5 mm for the scavenger stage (the geometrical slot pattern — actually a manufacturer´s demo sample — is shown in Figure 3). It is widely known that the selection of the grooved ferromagnetic matrix in this equipment impacts the separation efficiency. A Matrix gap which is very close to grain size makes it difficult to separate due to the particles’ mechanical entrapment inside the slots. Figure 3. Mini sample of matrix (for display purposes) in stainless steel corrugated plates Source: Prepared by the authors. The flow rate of slurry feed was 0.000294 m³/s (1.06 m³/h), and mass solid concentration was 40 %. The carousel angular velocity was 0.083 Hz (5.0 rpm). The washing water pressure was 41.0 kPa and electric current flowing through the electromagnetic coil was kept at 7.0 amperes. For the subsequent equipment sizing stage of the project, it is recommended, in principle, to use Equation 3, developed for carousel type high gradient magnetic separator, preconized by Silva and Luz (2013), for iron ores and Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 12 carousel diameter less than 4.0 m. 0.48 𝑑𝑐𝑎𝑟 0.498 𝑄𝑆 = 380 × [ ] × 𝑎𝑚𝑎 (4 − 𝑑𝑐𝑎𝑟 ) (3) Where: C = mass recovery (fraction of the total feed weight that reports to the concentrate) [–]; Qs = solid mass flow rate in feed stream [kg/s]; dcar = carousel diameter [m]; ama = is the matrix gap [m]. Magnetite scalping is particularly crucial in industrial processing routes when its content exceeds the 3% threshold. This is due to its high remanent magnetism, which can lead to accumulation of magnetic clusters in matrix gaps. While remediation strategies, such as increasing gap sizes or using matrices specifically designed for high flow rates (Ribeiro et al., 2017; Rocha et al., 2021), can mitigate this issue, it is important to note that these solutions must still achieve sufficient magnetic field gradients. Larger gaps can be effective as long as the required gradients are maintained. It is essential to remember that the effective separation mechanism relies on the magnetic field gradient, not solely on the magnetic field strength itself. On this issue, in the present work, visual inspection during the course of the experiments (concern driven by the 4.95 % magnetite content expected for a maximum feed content of 43.3 % Fe) showed that there was no need for prior removal of magnetite from the separator feed (at least for the mass processed in each experiment). Another magnetic separation study was carried out at Nomos Análises Minerais Company (at its Rio de Janeiro based lab), using a lab-scale wet high gradient magnetic separator Minimag®, also from Gaustec (Figure 4). A representative sample was firstly homogenized and quartered into three different size fractions, and then each subsample was comminuted in a disc mill, the first one up to 100 % below 212 µm, the second one below 150 µm and, finally, the third aliquot below 74 µm. The slurry was also fed at 40 % solids in mass. Figure 4 shows the equipment employed. Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 13 Figure 4. Wet high gradient magnetic separator (carousel type, vertical axis) Source: Prepared by the authors. The magnetic concentrate was collected, and the non-magnetic was fed again in the separator with higher magnetic field intensity (subsequent stage). This procedure was repeated until the seventh stage. Different magnetic fields — actually resulting in different magnetic gradients — were used in each stage according to Table 1. The assays achieved in each stage enabled calculations of what would be the grade if the initial sample were fed directly in each magnetic field. Stage 1 2 3 4 Table 1. Magnetic field (in teslas) used in each stage Magnetic field intensity [T] Stage Magnetic field intensity [T] 0.08 5 0.75 0.16 6 1.20 0.36 7 1.60 0.50 Source: Prepared by the authors. 3.3.4 Froth flotation The sample tested was the one below 150 µm. The pH adjustments were made with HCl (or H3C2OOH in a few cases) and NaOH. Cationic and anionic flotation techniques were tested. As anionic direct flotation is concerned, (semi) batch rougher flotation tests were carried out with sodium oleate and sodium sulfonate as collectors and ethanoic (acetic) acid and water glass as depressant. As sometimes kerosene or fuel oil is industrially used to improve collector efficiency, acting as chain extender, kerosene was additionally tested in this experimental campaign (Luz, 2015). Regarding reverse flotation, firstly two amines were compared — Flotigam Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 14 EDA 3 and Flotigam 2835 (both from Clariant) — using starch as depressant of a magnetic product from pilot magnetic separation held at Gaustec Company. According to Cassola and Bartalini (2010), Flotigam EDA 3 is an akylethermonoamine with about 30 % of neutralization by ethanoic acid. Flotigam 2835 is an akyl-etherdiamine. In reverse flotation tests, 80 g/t of collector (amine) and 500 g/t of depressant at pH 10.5 were used. Three depressants were tested, namely: corn starch, Brazilian tapioca starch, Angolan cassava starch. These tests were done using material below 150 µm, with or without desliming. Desliming, when adopted, was carried out using a cylindrical vessel of 0.23 m in height, removing particles below 10 µm by settling and subsequent siphoning (calculating the cut size by Stokes equation, as preconized by Braga, da Luz and Milhomem, 2018). The sample slurry was stirred and then was allowed to settle for 6 minutes. After the decantation step, the supernatant was immediately extracted by siphoning. The procedure was repeated four times (after water makeup addition). 4 RESULTS AND DISCUSSIONS 4.1 CHEMICAL CHARACTERIZATION STUDIES Chemical analysis by inductively coupled plasma spectroscopy (ICP) of the head sample (after standard chemical digestion) is shown in Table 2. Sieving results and iron distribution by size fractions is shown in Table 3. Table 2. Elemental analysis of the head sample Element Ca Fe K Mg Mn Na P Ti Content [%] 0.09 42.00 0.013 0.024 0.059 0.021 0.031 0.007 Source: Prepared by the authors. Size classes [µm] +1.200 - 1.200 + 800 - 800 + 600 - 600 + 425 - 425 + 300 - 300 + 212 - 212 + 150 - 150 + 104 - 104 + 74 Table 3. Sequence of formation of titles Retained [%] Iron grade [%] 2.67 22.11 12.31 8.23 8.89 4.76 3.44 4.07 3.51 43.20 40.65 49.15 44.52 44.91 48.62 47.56 44.58 41.41 Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 Iron distribution [%] 2.7 21.4 14.4 8.7 9.5 5.5 3.9 4.3 3.5 15 - 74 + 53 - 53 + 45 - 45 + 37 - 37 + 0 Total 3.88 2.76 0.55 22.81 100.00 38.55 37.59 34.50 35.89 41.98 3.6 2.5 0.5 19.5 100.0 Source: Prepared by the authors. A good correlation was found between iron grades by pycnometry and chemical analysis (Figure 5). Figure 5. Correlation between iron grade by pycnometry and chemical analysis Source: Prepared by the authors. Thermogravimetric analysis, differential thermal analysis plus Mössbauer spectrometry enabled the proportion quantification of the iron-bearing minerals: hematite (90 %); magnetite (8 %) and goethite (2 %). Mössbauer spectrum and sample thermogram are shown in Figure 6. The endothermic peak at 579.1 K (306 oC) is related to the transformation of goethite into hematite, as Ferreira and coworkers (2012) have already pointed out. In turn, the mass increase occurred around 1,073.1 K (800 °C) was assigned to the oxidation of magnetite to hematite, wherein the relative mass gain is about 3.5 %. In agreement with these results, X-ray diffractometry has revealed only the characteristic peaks of quartz (SiO2) hematite (Fe2O3) and magnetite (Fe3O4). Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 16 Figure 6. Spectra for the head sample. Upper: Mössbauer spectrum; Lower: thermogram Source: Prepared by the authors. 4.2 METALLURGICAL TESTWORK 4.2.1 Jig concentration The best Denver jig results (shown in Table 4) have revealed that iron ore concentrate was effectively separated in the size range between 0.6 mm and 1.0 mm (employing hutch water surface velocity Jf = 0.0156 m³/(s.m²) = 0.0156 m/s). Good grades were achieved in these conditions. Theoretically, separation would be feasible, since Taggart’s concentration criterion is 2.48, indicating that separation is possible up to 150 µm (disregarding particle shape effects preconized by Burt, 1984). Table 4. Bench scale jigging results Feed [% Fe] 50.53 50.53 50.53 Ragging thickness [mm] 80 175 245 Concentrate Tailing Gaudin´s selectivity index: Iron Recovery Iron Recovery 2 (1 − 𝑡) 𝑐 content, c [%] content, t [%] 𝑆𝐼 = √[ ]∙[ ] (1 − 𝑐) 𝑡 [%] [%] 64.97 7.00 47.76 93.00 1.42 62.86 50.97 39.00 49.03 1.63 50.50 99.54 38.85 0.46 1.27 Source: Prepared by the authors. Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 17 It can be inferred that the separation efficiency is higher in case of intermediate ragging thickness (175 mm). Thick ragging causes more friction during the suction cycle, hindering the fine concentration. In turn, excessively thin ragging allows detrimental passage of gangue particles, besides the denser particles (hematite, magnetite and goethite) through the hutch screen during suction. Therefore, ragging thickness of 175 mm showed the best efficiency (E = 44.9 %). On the other hand, separation efficiency of the others ragging thicknesses were E = -0.2 % for the higher level (negative value by reasons of analytical imprecision), and E = 7.2 % for the lower level. The corresponding results for Nomos’s jigging apparatus are shown in Table 5. The jig tailings contain around 15.0 % Fe, which can be, in a simplified approach, easily rejected without carrying out any further treatment, according to recent global practice. Therefore, jigging can be used as pre-concentration step for this iron ore. Table 5. Results using pilot hand jig Parameter Mass [kg] Feed 1.27 Concentrate 6.03 Middling 3.47 Tailing 1.77 Recovery [%] 100.0 53.5 30.8 15.7 Grade [% Fe] 3.59 55.44 7.24 15.67 Distribution [%] 00.00 68.05 26.30 5.65 SI 2.59 SI is the Gaudin´s selectivity index: 2 𝑆𝐼 = √[ (1 − 𝑡) 𝑐 ]∙[ ] (1 − 𝑐) 𝑡 Source: Prepared by the authors. It is noteworthy that the laboratory jig performed better than mini pilot jig, as expected, once lab experiments were performed under better control of variables, despite the use of jig with a much smaller jigging chamber area. 4.2.2 Humphrey´s spiral concentration The results of helix (“spiral”) concentration are shown in Table 6. Solid content and feed flow rate have influenced the separation efficiency. The experiment corresponding to the best separation efficiency was that one with 20 % Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 18 of solids by mass and slurry feed flow rate of 2.108 kg/s (7.59 t/h). Table 6. Results of spiral tests and its performance Stream Parameter Mass solid concentration Flow rate Iron assay Iron assay (c) Iron distribution Iron assay (t) Iron distribution Feed Concentrate Tailing 𝟐 Gaudin´s selectivity index 𝑺𝑰 = √[ (𝟏 − 𝒕) 𝒄 ]∙[ ] (𝟏 − 𝒄) 𝒕 Unit [%] [t/h] [% Fe] [% Fe] [%] [% Fe] [%] Mass solid concentration (cm) 20 25 30 7.59 5.44 6.91 46.25 46.25 46.25 51.46 52.00 49.44 91.10 77.20 87.00 22.67 33.67 32.31 8.90 22.80 13.00 [–] 1.90 1.46 1.43 Source: Prepared by the authors. 4.2.3 High gradient wet magnetic separation Magnetic separation was done on a pilot scale, displaying how the particle size has influenced the recovery of the ore minerals. The concentrate grades are presented in Table 7, while iron recoveries are presented in Table 7, for each size range (for the sake of brevity, the feed was designated as the "concentrate" under a null magnetic field.). Table 7. Actual grades achieved with magnetic separation Concentrate [% Fe] Magnetic Field Intensity [T] -212 µm + 150 µm -150 µm + 74 µm Feed 0.00 39.72 42.57 1 (rougher) 0.08 65.41 60.82 2 (1st scavenger) 0.16 61.84 60.54 3 (2nd scavenger) 0.36 58.95 59.88 4 (3rd scavenger) 0.50 58.09 59.28 5 (4th scavenger) 0.75 52.88 56.53 6 (5th scavenger) 1.20 52.01 55.73 7 (6th scavenger) 1.60 49.50 51.91 Final Tailing 8.27 7.41 Step -74 µm 43.32 66.70 64.76 64.16 63.72 59.66 57.93 53.34 8.33 Source: Prepared by the authors. The aliquot sized below 74 µm presented the best concentrates due to high assay in feed. However, lower recoveries are achieved in this class, indicating lower selectivity for fine grain size. The same effect has occurred in the laboratory scale experiment, probably due to the high liberation degree of the ore mineral (causing better sorting action) and to the lower probability of fine entrainment in the matrix (due to low particle Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 19 momentum of fines and consequently greater relative importance of the drag). From data of Table 7, one can simulate a rougher step performance for a given magnetic field intensity to be employed on an industrial scale, depending on particle size. Table 8 summarizes this reasoning, showing the expected performance, under several magnetic field intensities (in teslas). Table 8. Simulated iron grades and recoveries for versus magnetic field intensity Simulated performance Magnetic Between 150 µm and Between 74 µm and Below 74 µm field 212 µm 150 µm intensity Grade Recovery Grade Recovery Grade Recovery [T] [%] [%] [%] [%] [%] [%] 0.08 65.41 8.2 60.82 6.9 66.70 4.1 0.16 62.60 38.5 60.61 29.6 65.37 13.1 0.36 61.51 55.0 60.31 50.2 64.59 37.1 0.50 61.23 59.8 60.17 57.8 64.23 62.5 0.75 58.71 85.6 59.15 80.4 63.07 83.9 1.20 58.47 88.9 59.00 84.1 62.61 92.2 1.60 57.88 95.1 58.09 96.4 62.27 95.7 Source: Prepared by the authors. Using nonlinear regression for grade and recovery, it is possible to estimate these operational parameters in a single rougher industrial operation under a magnetic field intensity of 0.9 T, a set point very common in wet iron ore processing. After initial prospection, the expressions adopted were Equations 4 and 5: ✓ For concentrate iron grade: 𝑐 = 𝑒 −𝑘1 ∙𝐻 𝑚 (4) ✓ For iron recovery: 𝑅𝐹𝑒 = 𝑒 −𝑘2 ∙𝐻 𝑛 (5) The regression parameters for such equations can be seen ahead (Table 9). For the sake of concision, only the middle class (between 74 µm and 150 µm) plot of grade and recovery against magnetic field intensity is presented in Figure 7 (the other size classes’ behavior is very similar, as one can check from data displayed at Table 8 and 9). Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 20 Table 9. Regression for magnetic separation: concentrate grade and recovery and their coefficient of determination (R2) for the three size-class Parameter k1 k2 m n R2 for Equation 4 R2 for Equation 5 Regression parameters for size class between -150 + 212 µm +74 µm -150 µm - 74 µm 0.53039762 0.52709341 0.46328602 2.0979571 1.8915616 2.8170054 0.081507491 0.028127628 0.051909277 0.96468345 1.0174166 1.6513977 0.95406348 0.85128848 0.98341758 0.96133799 0.97924591 0.99154433 Source: Prepared by the authors. Of course, in virtue of the slightly high expected magnetite content (4.95 %, for 43.3 % iron feed content), a roll separator of low or intermediate intensity (rare earth permanent magnet roll separator) must be placed upstream, for an industrial circuit, in order to avoid clogging of the high gradient carousel-type separator. Figure 7. Simulated performance for Cateruca ore magnetic separation (size class between 74 µm and 150 µm) Source: Prepared by the authors. In sequence, the sizing of the high gradient separator can be done using Equation 3. Table 10 sums up the forecasting results for a hypothetical industrial rougher stage (disregarding a previous step for recovery of magnetite), using the regression equations earlier discussed. Table 10. Forecast for rougher operation at magnetic field strength of 0.9 T Size range: Iron content at feed (f): Iron content at concentrate (c): Iron content at tailing (t): Iron recovery (RFe): Gaudin´s selectivity index (SI): +150 µm - 212 µm +74 µm -150 µm 39.72 % 42.57 % 59.10 % 59.12 % 13.92 % 18.91 % 84.97 % 81.72 % 2.99 2.49 Source: Prepared by the authors. Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 -74 µm 43.32 % 63.08 % 10.75 % 90.63 % 3.77 21 4.2.4 Froth flotation The first reverse cationic flotation experiments were done aiming at cleaning a pre-concentrate (51 % Fe) obtained from pilot magnetic separation test. Amine Flotigam EDA 3 resulted 64.19 % Fe in the concentrate, 47.83 % Fe in tailing and 93.50 % of iron recovery, while Flotigam 2835 resulted 60.44 % Fe in concentrate, 39.44 % Fe in tailing and iron recovery of 89.76 %. Obviously in an industrial circuit the high content tailings would be reprocessed. The best collector tested was Flotigam EDA 3 for reverse flotation route. In sequence, the results using deslimed (and non-deslimed) ore are shown in Table 11. Table 11. Cationic flotation of sample deslimed below 10 µm corresponding to a feed assay of 46 % Fe (figures in brackets stand for the non-deslimed feed sample with 42 % Fe) Fe [%] Recovery [%] Fe [%] Recovery [%] 62.52 (59.42) 75.72 (79.97) 20.65 (19.35) 15.73 (15.58) Selectivity index, SI [–] 2.53 (2.47) 60.16 (59.69) 74.41 (75.52) 24.13 (21.94) 18.66 (18.06) 2.18 (2.30) 59.59 (57.05) 77.51 (79.78) 20.51 (20.58) 14.31 (15.20) 2.39 (2.26) Concentrate Tailing Depressant Corn starch Brazilian tapioca starch Angolan cassava starch Source: Prepared by the authors. The best test was with amine Flotigam EDA 3 and corn starch, resulting iron concentrate of 62.52 % Fe, with recovery of 75.72 %. Figures inside parentheses in Table 11 correspond to the non-deslimed sample (assaying 42 % Fe). In turn, the direct anionic flotation prospective experiments are shown in Table 12. The best result for one step direct flotation was obtained with sodium oleate (500 g/t) and kerosene (500 g/t) as collectors, and water glass (500 g/t) and ethanoic acid (100 g/t) as depressant in pH 4.0. This condition has provided a concentrate with 57.34 % Fe and iron recovery of 67.47 %. Table 12. Anionic direct flotation results (deslimed feed sample assaying 46 % Fe) pH [-] Ethanoic acid [g/t] Water glass [g/t] Sodium oleate [g/t] Sodium sulfonate [g/t] Kerosene [g/t] Concentrate content [% Fe] Metallurgical recovery [%] Gaudin´s selectivity index: [–] 8.0 6.0 4.0 100 0.00 200 500 650 500 600 600 350 0.00 300 350 600 0.00 400 47.56 43.94 48.54 84.55 83.87 80.34 1.19 0.71 1.24 Source: Prepared by the authors. Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 4.0 100 500 500 0.00 500 57.34 67.47 1.67 22 5 CONCLUSION Cateruca iron ore sample revealed a composition of 42.0 % Fe and 0.031 % P. X-ray diffraction identified quartz, hematite, and magnetite as the primary mineral phases, while Mössbauer spectroscopy quantified the ironbearing minerals as 90 % hematite, 8 % magnetite, and 2 % goethite. Analysis by Gaudin’s selectivity index indicated the best results for magnetic separation were achieved at a field of 0.9 T with particles below 74 µm. Jigging outperformed spiral concentration, yielding a concentrate with 62.86 % Fe and an iron recovery of 50.97 % when using intermediate ragging thickness. Spiral achieved its best results at 20 % solid content, producing a concentrate with 51.46 % Fe and an iron recovery of 62.6 %. High-gradient magnetic separation for Cateruca ore was efficient for material below 150 µm, with a recommended magnetic field strength of 0.9 T. Cationic flotation showed superior performance compared to the anionic route, although the expected results were not fully attained in the rougher stage alone. Froth flotation of the magnetic product, using mono-etheramine Flotigam EDA 3 (80 g/t) as a silicate collector and corn starch (500 g/t) as a depressant, achieved an iron grade of 64.19 % and a recovery of 93.5 % in the rougher stage. Corn starch was identified as the most effective depressant, yielding a concentrate with 62.52 % Fe and a recovery of 75.72 % for deslimed samples. Removing particles below 10 µm increased iron content in the concentrate but reduced mass recovery. Anionic flotation, using sodium oleate and kerosene (500 g/t) as collectors, with water glass (500 g/t) and ethanoic acid (100 g/t) at pH 4.0, produced concentrate with 57.34 % Fe and rougher recovery of 67.47 %. In the light of these results, the adoption of sequential concentration steps can be very beneficial. A hybrid route could be proposed, such as initial sorting by gravity methods and fine fraction concentration by magnetic separation or flotation in a subsequent step. Such a route would be cheaper, due to the partition of sized streams toward different process lines, since fine fraction processing is more expensive and complex. From an economic perspective, jigging and spiral concentration are more cost-effective due to their reliance on coarser particles and the absence of chemical reagents. In contrast, magnetic separation involves high Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 23 capital costs, and flotation incurs significant operational expenses, primarily due to reagent consumption. Future work should focus on scaling up a pilot plant to further investigate the influence of process parameters on mineral separability. This should include a trade-off analysis of different processing alternatives to optimize both technical and economic outcomes. ACKNOWLEDGEMENTS The authors thank Ferrangol, Prof. Geraldo Magela da Costa (Chemical Department/UFOP), geologist Mauro Yamamoto, Gaustec Magnetismo, and the UFOP’s Geochemistry Lab staff. The authors are thankful to Brazilian Council for Technological and Scientific Development (CNPq), Foundation for Research Support of the State of Minas Gerais (FAPEMIG) and Brazilian Federal Agency for Support and Evaluation of Graduate Education (CAPES) for their financial support. Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 24 REFERENCES APLAN, F. F. Gravity concentration (Chapter 6). In: FUERSTENAU, M. C.; HAN, K. N. Principles of mineral processing. Littleton: SME, 2003. ISBN 0-87335167-3. ARENARE, D. S.; ARAUJO, A. C.; VIANA, P. R. M.; RODRIGUES, O. M. S. Revisiting spiral concentration as applied to iron ore concentration. In: 2º International Symposium on Iron Ore, 2008, São Paulo, SP, Brazil. BRAGA, F. Y. A.; LUZ, J. A. M. da; MILHOMEM, F. de O. Size-dependent kinetics of iron ore reverse flotation. Particulate Science and Technology, v. 38, n. 4, p. 1-9, 2019. BURT, R. O. Gravity Concentration Technology: Developments in Mineral Processing. Amsterdam: Elsevier, 1984. CASSOLA, M. S.; BARTALINI, N. M. Behavior of collectors for the treatment of iron ore through reverse flotation. Tecnologia em Metalurgia, Materiais e Mineração, v. 6, p. 215-219, 2010. COUTO, H. J. B.; BRAGA, P. F. A.; FRANÇA, S. C. A. Use of gas pycnometry for estimating the iron content in mineral samples. Minerals Engineering, v. 39, p. 45-47, 2012. DAS, B.; PRAKASH, S.; DAS, S. K.; REDDY, P. S. R. Effective beneficiation of low-grade iron ore through jigging operation. Journal of Minerals and Materials Characterization and Engineering, v. 7, p. 27-37, 2007. DIEUDONNÉ, V.; JONKERS, A.; LOVEDAY, G. An approach to confidently predicting jigging performance. The Southern African Institute of Mining and Metallurgy, v. 106, p. 733-739. DOBBINS, M.; DUNN, P.; SHERELL, I. Recent advances in magnetic separator designs and applications. In: 7º International Heavy Minerals Conference 'What next', 2009, The Southern African Institute of Mining and Metallurgy. DWORZANOWSKI, M. Maximizing the recovery of fine iron ore using magnetic separation. The Journal of the Southern African Institute of Mining and Metallurgy, v. 112, p. 197-202, 2012. FERREIRA, M. M.; VARAJÃO, A. F. D.; MORALES-CARRERA, A. M.; PERALTASÁNCHEZ, M. G.; COSTA, G. M. Transformações mineralógicas e cristaloquímicas decorrentes dos ensaios termais em argilas cauliníticas. Cerâmica, v. 58, p. 105-110, 2012. GOVAERT, A.; DAUWE, C.; PLINKE, P.; DE GRAVE, E.; DE SITTER, J. A classification of goethite minerals based on the Mössbauer behaviour. Journal de Physique Colloques, v. 37, p. 825-827, 1976. KARON, S. Easy Plot --- Scientific Graphing and Data Analysis. Available at: http://www.spiralsoftware.com/EPManual.pdf. Access: 2014. Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 25 LUZ, J. A. M. Flotation of Iron Ore. In: GEORGE, E.; TOTTEN, R. C. Encyclopedia of Iron, Steel, and their Alloys. Bosa Roca: Taylor and Francis Group, 2015. MCKEE. Kassinga BHQ iron ore project. Cleveland, Ohio, 1974. OLUBAMBI, P. A.; POTGIETER, J. H. Effectiveness of gravity concentration for the beneficiation of Itakpe (Nigeria) iron ore achieved through jigging operation. Journal of Minerals and Materials Characterization and Engineering, v. 4, p. 21-30, 2005. RIBEIRO, J. P. et al. The challenge to scavenge iron from tailings produced by flotation a new approach: the Super-WHIMS & the BigFLUX magnetic matrix. REM: International Engineering Journal, v. 70, n. 3, 2017. ROCHA, R. B. da; REIS, E. L.; RIBEIRO, J. P. Wet high-intensity magnetic separators (WHIMS) for recovering iron from tailings deposited in dams. Mineral Processing and Extractive Metallurgy Review, v. 42, n. 1, p. 17-23, 2021. DOI: 10.1080/08827508.2019.1672061. SÃO JOSÉ, F.; BARCELOS, H. O.; PEREIRA, C. A. Combination of gravity concentration variables to increase the productivity of the Brucutu mineral processing plant. Journal of Materials Research and Technology, v. 7, p. 158164, 2017. SILVA, A. C.; TOMAZ, R. S.; SOUSA, D. N.; SCHONS, E. M.; BARROS, M. R. Influence of pulsation frequency in iron oxide jigging. In: 2º World Congress on Mechanical, Chemical, and Material Engineering, 2016, Budapest, Hungary. DOI: 10.11159/mmme16. SILVA, C. C. A concentração de minério de ferro com alto grau de hidratação. In: 2º Iron Ore Brazilian Symposium, 1999, Ouro Preto, MG, Brazil. SILVA, M. B.; LUZ, J. A. M. Magnetic scavenging of ultrafine hematite from itabirites. Rem: Revista Escola de Minas, v. 66, p. 499-505, 2013. SINUÉS, J.; SUBÍAS, I.; FANLO, I.; MANUEL, J. Geology and mineralogy of the Cateruca BIF deposit (Huila province, Angola). In: 11º Biennial Meeting SGA, 2011, Antofagasta, Chile. SIVAMOHAN, R. The problem of recovering very fine particles in mineral processing --- A Review. International Journal of Mineral Processing, v. 28, p. 247-288, 1990. TAGGART, A. F. Handbook of mineral dressing: ores and industrial minerals. New York: John Wiley, 1945. Studies in Engineering and Exact Sciences, Curitiba, Studies in Engineering and Exact Sciences, Curitiba, v.6, n.1, p.01-26, 2025 26